The proposed research investigates the neural bases of category learning in rule-based (RB) and information-integration (II) categorization tasks, and the extent to which learning in these tasks is based upon a procedural learning mechanism. The studies test two broad predictions of a biologically-plausible, multiple systems model of category learning (COVIS; Ashby et al., 1998). In the COVIS framework, category learning is hypothesized to be a competition between separate explicit and implicit systems. The implicit system is procedural learning-based and assumed to mediate learning in II tasks. In contrast, the explicit system is a logical-reasoning system that is assumed to mediate learning in RB tasks. Learning in the explicit system is assumed to rely primarily upon frontal cortical structures and the head of the caudate nucleus whereas the implicit system depends primarily upon the tail of the caudate nucleus and high-level motor structures. The first two experiments test the hypothesis that learning in the implicit system is procedural learning-based and strongly tied to motor systems whereas learning in the explicit system is RB and more abstract. The remaining four experiments use neuropsychological and neuroimaging techniques to test the hypothesis that the neural systems mediating performance in RB and II tasks are dissociable. The experiments also entail a novel neuropsychological direction in category learning research in investigating the role of the cerebellum in these tasks. This research is important because it uses a converging operations approach to provide insight into the learning algorithms and neural mechanisms involved in RB and II category learning tasks.